signup-flow-cro
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides purely instructional content for conversion rate optimization. No malicious patterns, obfuscation, or unauthorized access attempts were identified.
- [INDIRECT_PROMPT_INJECTION]: The skill instructions involve reading local context files, which creates a potential ingestion surface for untrusted data.
- Ingestion points: The skill reads
.agents/product-marketing-context.mdor.claude/product-marketing-context.mdas specified inSKILL.md. - Boundary markers: None specified for the ingested file content.
- Capability inventory: No dangerous tools, network operations, or shell commands are defined in the skill.
- Sanitization: No sanitization logic is present for the ingested context.
Audit Metadata